Report
GitHub - openai/openai-cookbook: Examples and guides for using the OpenAI API
Examples and guides for using the OpenAI API. Contribute to openai/openai-cookbook development by creating an account on GitHub.
github.com
I found OpenAi Cookbook pretty useful. Do you know other useful GPT-3 resources?
This blog post about building on top of OpenAI's quickstart template was pretty interesting: https://www.propelauth.com/post/production-ready-openai-gpt-template
Thanks @scrappyoperator it is very interesting
Please check https://github.com/openai/openai-cookbook/blob/main/techniques_to_improve_reliability.md
Thanks @silviwillemson I will have a look :)
Yes!
Thanks for the reply @davidalex
Hey Ramón, YES! After seeing this post and countless others looking for prompt resources, we launched an app called Promptly on ProductHunt today! Promptly is a free tool to share, discover & save your favorite AI prompts. We support the biggest services (ChatGPT, Midjourney etc.) and want to make it as easy and fast as possible for users to hunt down cool new AI use cases.
You can check it out at https://promptly.cc :)
Really looking forward to hearing what you think, and also looking forward to implementing feedback made in here!
Best, Mateo
Nice one! Just had a look at it and love the design 😍
I will use it for sure in the future, however, some prompts are not that "complex"
For example, there was a "Create A LinkedIn Post" which prompt was "Create a standard CEO post on LinkedIn" 😅
Thanks, nice reference.
I like the open-sourced alternative to GPT-3: GPT-J
https://developer.nvidia.com/blog/deploying-gpt-j-and-t5-with-fastertransformer-and-triton-inference-server/
Thanks for sharing that article @terry_remyx 😄
Looks like you have to deploy it yourself. Have you tried doing so? Is it worth it?
I have and there are some details to work out but here is a docker image: https://hub.docker.com/repository/docker/smellslikeml/triton_with_ft
For some applications, serving GPT-J can be a more cost effective than using an API.
NVIDIA's tritonserver uses kernel autotuning to optimize inference for the specific hardware setup used in deployment.
Love the idea behind this. Thanks for the extra info. I will save this and have a look if the OpenAI costs escalate too much on my current app. Thanks!
Thanks @reymon359 I am currently building https://ai.engage.ninja/ and this cookbook is really helpful. Bookmarked!
Glad you liked it!
Yes there many resources but Here few Resources
ChatGPT Prompts : https://chatgptconnect.com/chatgpt-prompts/
ChatGPT Detector : https://chatgptconnect.com/chat-gpt-detector/
Chatgpt Chrome Extension: https://chatgptconnect.com/chatgpt-chrome-extensions/
Hope you find it useful
for complete resource check this : https://chatgptconnect.com/
Embeded chatGPT on Google broswer: https://chrome.google.com/webstore/detail/chatgpt-for-google-chrome/clegjkopllnjgmbgeldhbebmjgghcdci
Give it a try and write down your feedback guys..🙂
https://gpt3sandbox.com/
https://gpt3examples.com/
Something that doesn't pop up right away in the guide but has given me so much insight about how GPT-3 works and prompt design, is this part of the cook book: https://github.com/openai/openai-cookbook/blob/main/techniques_to_improve_reliability.md
Basically how to improve reliability of your prompt and how you can design your prompt to make it more understandable for GPT-3
Thanks for sharing that link @hzzhdev, the analysis from there is gold!
I recently released prompto, a simple directory of community submitted prompts.
👉 https://prompto.chat/
Wow, looks great. What is the main difference between this one and others like promptly.cc ? 🤔
Promptly beat us to it by a couple of days 😅 but we are hoping to build a community around AI and it's use cases, we just added community posts for AI-focused discussion and have plans for more community focused features.
No worries, I really liked your search interface more than theirs to be honest. Like I could find more useful things in yours. Keep up the good work!
Super cool! This isn't released yet but might be potentially useful to see exactly where ChatGPT is getting its sentence info from
I've recently launched my product www.habitcoach.ai and found David Shapiro's videos on youtube to be immensely helpful in wrapping my head around how to build something with GPT-3.
All his videos are at: https://www.youtube.com/c/DavidShapiroAutomator
Thanks a lot @singularityguy for the tip!
Thank you pro 🤍
I really liked the documentation on OpenAI's website. I found it to be really thorough.
Thanks for answering, and I 100% agree. Their docs were super helpful when I started working on buildpublic.app :)
Hi Ramon,
We recently built https://ansy.ai/, it is a Discord bot that will answer community questions based on chat history of public channels. Cool way to save time of mods. We launched it on Product Hunt as well. https://www.producthunt.com/posts/ansy-ai . Do check us out and let us know what you think.Your feedback will be super helpful.
Thank you for sharing your project with me, Ansy.ai sounds like a useful tool for Discord communities. Based on your description, it seems that Ansy.ai uses chat history from public Discord channels to provide answers to community questions.
Does Ansy.ai support integration with Slack or other messaging platforms in addition to Discord? 😄
Hi yes, We're releasing for discord early Feb. And then we'll be releasing for WhatsApp.
You can sign up to our waitlist and you'll get notified immediately
Cool, thanks for clarifying!
This is interesting, may be there a need for a tool which helps in writing promts to chatgpt with a proper prompt engineering, for deriving the best results out.
Awesome Idea @NileshBhosale, I bet sooner or later it will appear indeed